Skip to content

Guides

Describe a piece of software, and the work is split into parts that stand alone, built in parallel, and each part checked by something that did not write it. On your hardware, against the models you choose.

These guides cover the machinery that runs that process: standing up the platform, and configuring the agents, roles, and approval gate that carry out and check the work. Each guide is self-contained with a clear goal.

Pre-alpha

SynthOrg is pre-alpha. The loop has been driven live against a real deployment twelve times and has never reached the assembly stage: no run has produced an assembled deliverable, and no completion has been recorded. Nothing on these pages promises you working software; they describe how the system is built and how it behaves, which you can check against the code. See the Roadmap for what is wired versus what is intent.

New to SynthOrg?

Start with the Quickstart Tutorial to stand up the platform and configure a company in about 5 minutes.

  • Quickstart Tutorial

    Install the CLI, pick a template, and stand up the platform in about 5 minutes.

    Quickstart

  • Company Configuration

    Complete YAML reference for every configuration option.

    Configuration Reference

  • Agent Roles & Hierarchy

    Define agents, roles, departments, and reporting lines.

    Agents

  • Budget & Cost Control

    Per-agent budgets, alert thresholds, quota degradation, and spending reports.

    Budget

  • Security Policies

    Autonomy levels, approval gates, and custom security policies.

    Security

  • Tool Integration (MCP)

    Connect external tools via MCP servers with stdio or HTTP transports.

    MCP Tools

  • Deployment (Docker)

    Run SynthOrg in production with Docker, hardening, and image verification.

    Deployment

  • Memory Configuration

    Agent memory, shared org memory, retrieval pipeline, and consolidation.

    Memory

  • Centralised Logging

    Route structured logs to syslog, HTTP, OTLP, or external log aggregators.

    Centralised Logging

  • Settings Reference

    Resolve, view, and edit the 35 settings namespaces at runtime.

    Settings Reference

  • Notifications & Events

    Configure notification sinks and subscribe to WebSocket event channels.

    Notifications & Events

  • Workflow API Tutorial

    curl tutorials for creating, versioning, activating, and executing workflows.

    Workflow API

  • Agent Management

    Hire, fire, promote, and customise agents through the REST API.

    Agent Management

  • Human Interaction & API

    REST + WebSocket API surface, rate limiting, RFC 9457 errors, Web UI features.

    Human Interaction

  • Contributing

    Development workflow, testing, code style, and pull request process.

    Contributing

  • Adding a Provider

    Configure LLM providers, models, and connection secrets.

    Adding a Provider

  • Approval Workflow

    Approve or reject parked agent actions through the API and dashboard.

    Approval Workflow

  • A2A Federation

    Federate agents across organisations with the agent-to-agent protocol.

    A2A Federation

  • Cost Attribution

    Attribute spend to agents, projects, and tasks for chargeback.

    Cost Attribution

  • Custom MCP Server Development

    Build your own MCP server to expose bespoke tools to agents.

    Custom MCP Server

  • Custom Rules & Meta-Loop

    Author declarative signal rules that drive the self-improvement loop.

    Custom Rules

  • Dynamic Scoring

    Configure candidate ranking and model-selection scoring strategies.

    Dynamic Scoring

  • Fork Setup

    Stand up your own fork with CI environments and required secrets.

    Fork Setup

  • Monitoring & Dashboards

    Observe health, metrics, and live agent activity.

    Monitoring

  • Ontology Extension

    Define entity types and extend the semantic ontology.

    Ontology Extension

  • Persistence Migrations

    Manage schema revisions across the SQLite and Postgres backends.

    Persistence Migrations

  • End-to-End Run

    Run a real objective through the whole loop and read the result back from the dashboard.

    End-to-End Run

  • REST API Examples

    Copy-paste curl recipes for common REST API operations.

    REST API Examples

  • Webhook Management

    Register, verify, and observe inbound and outbound webhooks.

    Webhook Management

  • Workers & Background Tasks

    Run the worker pool and the background task queue.

    Workers & Background Tasks

  • Project Brain

    Query a project's structured decision log: decisions, open questions, blockers, risks, dependencies, and plan revisions.

    Project Brain

  • Living Documentation

    Browse and search the per-project document wiki agents write to and retrieve from automatically.

    Living Documentation

  • Knowledge Substrate

    Ingest external corpora (PDFs, web pages, repositories) into a provenance-tracked retrieval layer agents can cite.

    Knowledge Substrate

  • :material-rudder:{ .lg .middle } Mid-Flight Steering

    Inject hints and redirects into a running project and supersede obsolete tasks without stopping it.

    Mid-Flight Steering

  • Research Mode

    Run recorded, replayable research briefs through a pipeline that synthesises cited reports.

    Research Mode

  • Brownfield Intake

    Import an existing codebase into a project, build a structure map, and run an automated analysis pass.

    Brownfield Intake